Wired has posted a new
article on the topic self-reconfiguring robots. It covers the basic
concept of self-reconfiguring robots and the current state of the art.
The article focuses primarily on macro-scale reconfiguration (e.g.
transformer-style combinations of several robots) but does touch on the
idea of nano and atomic scale reconfiguration.
...quite simply, it's often hard enough to program one bot to work or a
cluster of them to work loosely together. This seems an even greater
nightmare! I've been following the progress of this concept for a few
years now and, although the modules are coming on in leaps and bounds
(some of them will even self connect if placed near each other on the
table), they all seem to be pretty much hand built and hand coded to do
a task.
Role on ditributed, evolving systems :)